Jev models and aliases
Jev models and aliases are the ids you put in the request model field on POST https://api.typesafe.ai/v1/systemone. They are not a second evaluation URL. From the public models table:
| Name | Role |
|---|---|
jev-1.13.0 |
Versioned ID |
jev-latest |
Alias → current stable official release (jev-1.13.0 on the snapshot we read) |
jev-preview |
Alias → newest release, including previews; currently the same ID |
This page is for builders wiring TypeSafe Jev (System One) into agents, routing, or evals. jev.pro is an independent unofficial field guide — not affiliated with TypeSafe AI. We do not sell or issue keys. Confirm the live table on docs.typesafe.ai models and the API reference.
SDK examples use jev-latest. The response model field reports the versioned ID. GET /v1/models returns names, descriptions, and release dates. It currently lists aliases. You may still send versioned IDs even if they are absent from that list. Do not send truncated names such as jev-1.13 unless the models page says they exist.
Why pin
Aliases move when TypeSafe ships. If you tuned a 0.85 confidence gate on 1.13, a silent move can change the distribution. Pin jev-1.13.0 in production, upgrade in a controlled diff, and keep the resolved id in logs. Replay the same state + questions against the new pin before you flip traffic.
jev-preview is the “including previews” alias. Today it points at the same id as jev-latest on the snapshot we read. Treat that as confirm language, not a promise they stay coupled.
What does not change with the model string
The HTTP path stays /v1/systemone. Primitives stay Choice (max 255, unordered), Score (2–10 levels), and Noul (P(yes) 0–1, no confidence). Question ids are still not sent to the model. Text only; English primary. Documented errors stay 401 / 422 / 429 / 529. Official SDKs still retry 429/529.
Vendor claims on the models page
Label these as TypeSafe claims, not measurements from jev.pro: $0.042 / million input tokens, output free; 250k tok/s; 1200 RPM; 64k context for state + all questions; 32k for state + the longest question; launch-post 70ms–500ms / 40x–200x. Do not invent SLOs on top.
Gateway spelling
typesafe-ai/jev is a Gateway model id in Vercel docs, not the first-party model field. Sending it to api.typesafe.ai is a door-specific failure, not a documented alias.
See also
- API hub — evaluation POST and alias list GET
- Jev 1.13 — current public model notes
- Glossary: jev-latest — alias definition
- Version pinning — promotion checklist
FAQ
Is typesafe-ai/jev a TypeSafe alias?
That spelling is a Gateway model id in Vercel docs, not the first-party model field.
Preview vs latest today?
Same id on the snapshot we read. Confirm the models page.
List price on this table?
Vendor claim on the models page — pricing.
Can GET /v1/models replace a health check?
It is a cheap authenticated ping. It is not a job-status API and it does not list every versioned id you may POST.
What this page does not claim
- Official TypeSafe status, logos, or support.
- That jev.pro sells, issues, or proxies API keys.
- Extra first-party HTTP routes beyond
POST /v1/systemoneandGET /v1/modelsunless the API reference lists them. - Invented latency SLOs, search volume, or unofficial prices.
- That a schema-constrained answer is automatically factually correct.
Disclaimer
This is an independent unofficial site and is not affiliated with TypeSafe AI; official documentation is available at https://docs.typesafe.ai. Never treat jev.pro as TypeSafe official documentation. We do not sell, issue, or proxy API keys.